Archimago has a new DAC listening test

[herr.jemine]

Logic needs to be rte-examined when it goes against facts. I suggest you do that instead of arguing.

Quite simply, I think: the test setup influences the psychoacoustic perception of the listener at the end of the chain.
The audio signal is minimally changed by the conversion in the ADC and does not correspond to the original.
minimal timing / volume changes have an effect on the spatial perception of the sound stage, the perceived distance of objects in the room, etc.
It is not possible to evaluate the quality of DACs with this test.

 

[solderdude]

Why would it be possible to record music including the nice spatial effects using an ADC (even through various digital effects) but an ADC is not able to record the analog signal of 2 DACs with fidelity ?

 

[SIY]

Quite simply, I think: the test setup influences the psychoacoustic perception of the listener at the end of the chain.
The audio signal is minimally changed by the conversion in the ADC and does not correspond to the original.
minimal timing / volume changes have an effect on the spatial perception of the sound stage, the perceived distance of objects in the room, etc.
It is not possible to evaluate the quality of DACs with this test.

Try to be specific in your objections rather than hand-wavy. "I think" and "minimal" are not exactly part of analysis.

 

[ahofer]

Quite simply, I think: the test setup influences the psychoacoustic perception of the listener at the end of the chain.
The audio signal is minimally changed by the conversion in the ADC and does not correspond to the original.
minimal timing / volume changes have an effect on the spatial perception of the sound stage, the perceived distance of objects in the room, etc.
It is not possible to evaluate the quality of DACs with this test.

Please present evidence: a) the magnitude of the alleged phase changes and b)they are audible and salient as described.

Does Phase Distortion/Shift Matter in Audio? (no*)

A viewer of my videos and member he suggested that I do a video commenting on a video that Paul McGown did on audibility of phase shifts. Here is Paul's video which was really about a different question (why we need wideband amplifiers) but turned it into phase being an audible problem: Here...

audiosciencereview.com

 

[krabapple]

Quite simply, I think: the test setup influences the psychoacoustic perception of the listener at the end of the chain.
The audio signal is minimally changed by the conversion in the ADC and does not correspond to the original.
minimal timing / volume changes have an effect on the spatial perception of the sound stage, the perceived distance of objects in the room, etc.
It is not possible to evaluate the quality of DACs with this test.

Do tell us just how 'minimal' the changes need to be, to NOT have an audible effect. Surely there must be a threshold. Ears are not infinitely resolving. So give us a number and Tthen we compare your claim to known psychoacoustic science, and to the measuments of these devices.

And btw, your IRL name isn't 'Milind Kunchur' , is it?

 

[herr.jemine]

oh boy, no I won't do that - this is not an interrogation - keep living your religion and everything will be fine

 

[krabapple]

I'm asking you for data that supports your claim. What's the minimal change from ADC that humans can hear, according to you? Or do you assert there is no limit: any change is audible?

The Kunchur ref was a wink to folks here who've been down this road before, even if you apparently haven't.

 

[herr.jemine]

I just thought it was wrong to measure like that, I don't have any data... it's just an assertion - you can only make a comparison by listening to the ADC recordings in the listening room with the actual outputs of the DAC on the same listening station and comparing them.

why does an aural exciter work? The signal to be processed is split, the part to be processed is high pass filtered and distorted so that harmonic overtones of the signal are added. This part is then subtly mixed back into the original signal.
a minimal change but the psychoacoustic effect is enormous.

 

[ahofer]

I just thought it was wrong to measure like that, I don't have any data... it's just an assertion - you can only make a comparison by listening to the ADC recordings in the listening room with the actual outputs of the DAC on the same listening station and comparing them.

why does an aural exciter work? The signal to be processed is split, the part to be processed is high pass filtered and distorted so that harmonic overtones of the signal are added. This part is then subtly mixed back into the original signal.
a minimal change but the psychoacoustic effect is enormous.

But what does this have to do with the topic at hand? Surely you don’t think the ADC conversion is doing something equivalent to the Aural Exciter (which is trivial to measure).

It seems like the one “just making assertions” and arguments to mystery might be a little closer to a “religion” here.

 

[Ken Tajalli]

Again, the result is the comparison of outputs of multiple DACs and that was the goal. It produced a high statistical significance. You need a bit more than to make claims that it’s invalid, since the result shows the opposite. The effect of jitter or other distortions is known for this ADC and it’s far below the errors in the output of any of these DACs.

Consider this:
Almost all DACs do their worst with 16/44 materials i.e. CD quality.
DACs with 24 bit material, have an extremely low noise floor. Also with 44.1kHz materials, the reconstruction filters used could arguably have an audible effect.
Archi used 16/44 standard CD quality pieces of music, to give the DACs under test the opportunity to show their audible differences.
16/44 material technically has limits at the edge of audiblity.
Archie then recorded the outputs of the DACs using an state of the art ADC, at 24 bits and 96kHz. at these levels, the limits of the system is many times above audiblity, so can be accepted as Transparent. Also the ADC would be capable of capturing ultrasonic noises, the noise floors, that the test DACs may be outputting.
Also, on playback side, the various DACs used, should have very little effect on sound quality, since they, also, are being fed a 24/96 signal, which is much easier to handle.

 

[pkane]

Archi used 16/44 standard CD quality pieces of music

Did he?

Consider this:
Almost all DACs do their worst with 16/44 materials i.e. CD quality.
DACs with 24 bit material, have an extremely low noise floor. Also with 44.1kHz materials, the reconstruction filters used could arguably have an audible effect.
Archi used 16/44 standard CD quality pieces of music, to give the DACs under test the opportunity to show their audible differences.
16/44 material technically has limits at the edge of audiblity.
Archie then recorded the outputs of the DACs using an state of the art ADC, at 24 bits and 96kHz. at these levels, the limits of the system is many times above audiblity, so can be accepted as Transparent. Also the ADC would be capable of capturing ultrasonic noises, the noise floors, that the test DACs may be outputting.
Also, on playback side, the various DACs used, should have very little effect on sound quality, since they, also, are being fed a 24/96 signal, which is much easier to handle.

Ok, I considered Is there a conclusion I should draw from this?

 

[Ken Tajalli]

Did he?

Yes he did.

Edit: He may not have! It seems the original tracks were one at 24/88.2 (from 16/44 original) one at DSD (again from CD original), the rest 16/44.1.
But then it seems he used Hires resamples of the same tracks. I just looked it up, and frankly am a bit confused.

Ok, I considered Is there a conclusion I should draw from this?

Sure. Conclusion is, that I am right.

 

[pkane]

Yes he did.

Edit: He may not have! It seems the original tracks were one at 24/88.2 (from 16/44 original) one at DSD (again from CD original), the rest 16/44.1.
But then it seems he used Hires resamples of the same tracks. I just looked it up, and frankly am a bit confused.

I think Archi said he used 24 bit files for this survey. Whether or not the original source material was at 16 bit or not, I can't really tell...

Sure. Conclusion is, that I am right.

I've reached a different conclusion

 

[board]

yes, only electrical voltage was measured and recorded, that is my statement. through DAC-ADC-DAC what is the point? the signal is only contaminated - the ADC process more or less harmonizes the signal through its jitter and the way it records, a statement on this is not relevant, such tests can only be carried out on the object with headphones and speakers and not from a distance, that is nonsense!

You forgot to mention that vinyl is superior to digital, and that you don't need to take any kind of test to show that you can hear the effect of digitization, because it's so goddamn obvious that anybody could hear it - even your wife!

 

[Ken Tajalli]

I think Archi said he used 24 bit files for this survey. Whether or not the original source material was at 16 bit or not, I can't really tell...

These are the first three tracks.

did he use 16/44 from higher res material (A & C), or are they higher res files from 16/44 originals, I don't get it.
The other songs are standard 16/44.
However, elsewhere, he seems to indicate that he used 24 bit versions of the tracks! did he process them? Why or for what purpose ...
I don't know.

I've reached a different conclusion

If your different conclusion is not a personal one, then I am curious.

 

[herr.jemine]

You forgot to mention that vinyl is superior to digital, and that you don't need to take any kind of test to show that you can hear the effect of digitization, because it's so goddamn obvious that anybody could hear it - even your wife!

yes exactly, your mother sounds good too

 

[krabapple]

yes exactly, your mother sounds good too

you are so not getting it

 

[krabapple]

These are the first three tracks.
View attachment 372970
did he use 16/44 from higher res material (A & C), or are they higher res files from 16/44 originals, I don't get it.
The other songs are standard 16/44.
However, elsewhere, he seems to indicate that he used 24 bit versions of the tracks! did he process them? Why or for what purpose ...
I don't know.
View attachment 372972

If your different conclusion is not a personal one, then I am curious.

He did a different test in 2021, using (downconverted when necessary) 16/44 versions of his AMPT Test Tracks. That''s what you quote in your first image.

For this 2024 DAC test, he used/made 24 bit versions of the AMPT tracks. That's what you quote in your second image.

 

[danadam]

did he use 16/44 from higher res material (A & C), or are they higher res files from 16/44 originals, I don't get it.

Seems quite clear to me that they are hires downsampled to 16/44. Although I'm not sure where does the 24/88k version of Stravinsky come from. The Reference Recording site offers only 16/44k and 24/176k. Later in the post he writes:

the Stravinsky piece (this is not the HDCD version but sourced from 24/88.2 Reference Recordings hi-res)

He also writes there:

Let's stick with lossless 16/44.1 source. No need to argue or debate over hi-res these days I think (see here, and here). Even if you can hear a difference with hi-res, all indications point to the difference being minuscule.

I guess that for the blind test he left them at 24-bit to avoid complaints that 16-bit was the reason for not hearing differences.

 

[Ken Tajalli]

He did a different test in 2021, using (downconverted when necessary) 16/44 versions of his AMPT Test Tracks. That''s what you quote in your first image.

For this 2024 DAC test, he used/made 24 bit versions of the AMPT tracks. That's what you quote in your second image.

Sounds about right.
But why did he use a 24 bit version of a 16 bit track, on this occasion?

guess that for the blind test he left them at 24-bit to avoid complaints that 16-bit was the reason for not hearing differences.

But 16 bit original will remain 16 bits, even in a 24 bit envelope.